Quantifying Uncertainty: Prediction and Inverse Problems

Models often contain parameters which are not known exactly. We examine mathematical methods to both estimate parameters from data and to quantify the uncertainties in the outputs from the models.

Course leader

Laura Scarabosio
Assistant professor
Mathematics
Radboud University

Björn Sprungk
Assistant professor
Applied Mathematics
TU Bergakademie Freiberg

Target group

-PhD
-Post-doc
-Professional

The course targets PhDs students, postdocs and professionals who are eager to learn more about uncertainty quantification and Bayesian inverse problems, the possible algorithms that can be used depending on the specific mathematical model as well as their theoretical foundations.

Course aim

After this course you will be able to:

- Choose the best suited algorithm to perform uncertainty quantification for a specific problem.
- Use and implement Monte Carlo, multilevel Monte Carlo and stochastic collocation.
- Formulate a Bayesian inversion problem and study its well-posedness.
- Use and implement Markov chain Monte Carlo methods.

Fee info

EUR 325: The fee includes the registration fees, course materials, access to library and IT facilities, coffee/tea, lunch, and a number of social activities.

Scholarships

We offer several reduced fees:
€ 293 early bird discount- deadline 1 April 2022 (10%)
€ 276 partner + RU discount (15%)
€ 244 early bird + partner + RU discount (25%)